2.1 Canonical Pathway: SMAD-Dependent Signaling

The TGF-β signaling is triggered by phosphorylation and activation of its transmembrane receptors, which then deploy SMAD transcription factors to regulate critical cellular processes. These mechanisms have been comprehensively reviewed elsewhere [3, 9]. Briefly, the TGF-β ligands initially bind to the TGF-β Type III receptors (TGF-βRIII), which then present TGF-β to the TGF-βRI/TGF-βRII complex, which has a high affinity for TGF-β. Activated TGF-β receptors phosphorylate SMAD2/3 on two C-terminal serine residues (SSXS motif). Phosphorylated SMAD2/3 bind SMAD4 to form a heterotrimeric complex, which translocates to the nucleus to activate or restrain target gene expression [10-13]. This is called canonical TGF-β signaling. The biological effects of this signaling are context and tissue-dependent. It could induce growth [14], cell death [15], metastasis [16, 17], differentiation [18, 19], remodeling [20-22], and stress response [23-25] (Figure 2). The concept of context-dependent biological activities of TGF-β is further discussed in Section 3 of this article.

2.2 Non-Canonical Pathway: SMAD-Independent Signaling

Apart from this SMAD-mediated effect, TGF-β signaling can also mediate its biological effect in a SMAD-independent mechanism. This is called non-canonical TGF-β signaling, which occurs via pleiotropic pathways (Figure 3). Multiple SMAD-independent (non-canonical) signaling pathways, including the mitogen-activated protein kinase (MAPK) pathway, which comprises extracellular signal-regulated kinase (ERK), jun N-terminal kinase (JNK), and p38 MAPK, phosphatidylinositol 3-kinase/protein kinase B (PI3K/AKT) are activated in response to TGF-β [11, 12, 23-31]. The TGF-β-activated transforming-beta receptor-I (TβRI) can directly phosphorylate an adapter protein called Src homology domain 2-containing protein, forming a complex with growth factor receptor-bound protein 2 (GRB2) and son of sevenless (SOS) [30]. This cascade activates rat sarcoma (Ras), which sequentially phosphorylates rapidly accelerated fibrosarcoma, mitogen-activated protein kinase (MEK), and ERK, ultimately regulating cell survival, proliferation, adhesion, migration, and metabolism. ERK activation is implicated in multiple diseases, including cancer, making it a crucial pathway in TGF-β signaling.

TGF-β also triggers Rho guanosine triphosphate hydrolases (Rho GTPase) signaling to influence cytoskeletal dynamics and cell migration, usually through epithelial-to-mesenchymal transition. For instance, TGF-β activates Ras homolog gene family member A (RHOA), which activates its key effector Rho-associated protein kinase 1 (ROCK1), leading to the activation of Lin-11, Isi-1, and Mec-3 (LIM kinases) to phosphorylate cofilin, a constitutive actin-depolymerizing factor. This phosphorylation cascade ultimately results in the stabilization of the actin cytoskeleton [32, 33]. TGF-β activates other Rho GTPases, such as cell division control protein 42 (CDC42), to engage cytoskeletal rearrangement, EMT, and invasion [34].

Collectively, activation of Rho GTPases by TGF-β reprograms cell dynamics in a way that facilitates cell migration, invasion, and metastatic spread. The TβRI kinase activity can also activate the PI3K/AKT signaling pathway through the classical mechanism involving activation of the PI3K- AKT-MTOR-S6k axis. This signaling is prominent for promoting cancer metabolic reprogramming, cell survival, and proliferation [35].

Another critical signaling cascade activated by TGF-β is the Janus kinase-signal transducer and activator of transcription (JAK/STAT) signaling, particularly involving JAK1 and JAK2 in certain contexts. This includes hepatic stellate cells, where activated JAK phosphorylates STAT3 to mediate the fibrogenic effects of TGF-β signaling [36]. Beyond fibrosis, JAK/STAT signaling is also implicated in immune regulation, wound healing, and tumor progression [37], making it a key pathway in TGF-β-mediated physiological and pathological processes.

The TGF-β/SMAD signal transduction pathway is tightly regulated through positive and negative mechanisms at multiple stages. A key negative feedback mechanism involves SMAD7, an adaptor protein induced by TGF-β transcriptional activity, which recruits SMAD ubiquitination regulatory factor (SMURF) ubiquitin ligases to the receptor complex, leading to its degradation via proteasomal and lysosomal pathways (Figure 2). In addition, autophosphorylation of serine residues on Type II receptors results in positive and negative regulation [38]. For instance, Luo and Lodish, using mouse and human cell lines, showed that phosphorylation of Ser213 activates transforming-beta receptor II (TβRII) while phosphorylation of Ser416 inhibits its kinase activity [39]. The appropriate regulation of the TGF-β signaling pathway prevents congenital and acquired musculoskeletal disease [3, 40, 41], like the accumulation of extracellular matrix (ECM)—loss of its regulation results in impaired regeneration of both inherited and acquired neuromuscular disorders. Overexpression of TGF-β in mice has been reported to induce bronchopulmonary dysplasia [42, 43], hydrocephalus [44-46], and impaired skin [47]. In humans, overactivation of TGF-β results in heritable Camurati-Engelmann disease (CED), seen as hyperostosis, sclerosis of the bones [48, 49] and cancer. When TGF-β function is lost in humans due to mutations in TGF-β or its receptor, aortic aneurysms and skeletal fragility are seen [50-52]. These findings underscore the importance of tightly regulated TGF-β signaling in maintaining tissue homeostasis, which is essential for preventing various pathological conditions.

3.1 The Anti-Tumor Roles of TGF-β

The anti-tumorigenic role of TGF-β is predominantly highlighted in the early phases of cancer development, wherein it acts to suppress cancer cell proliferation, induce programmed cell death (apoptosis), ensure genomic stability, and augment immune surveillance. These functions are critical for maintaining cellular homeostasis and preventing the development of malignant cells. The subsequent section delineates this activity in greater detail.

3.2 The Pro-Tumor Roles of TGF-β

In malignant and advanced tumors, the overexpression of TGF-β cells is common in most tumor microenvironments (TME), leading to various tumor-promoting processes [22]. Increased oncogene expression, the release of immunosuppressive cytokines, and epithelial plasticity are all linked to TGF-β overexpression during tumorigenesis [74]. In the TME, stromal fibroblasts and other cancer cells release TGF-β, which shapes the tumor's architecture while suppressing the immune cells’ anti-tumor function, ultimately leading to tumor proliferation and progression [74]. TGF-β's pro-tumor activities result from its involvement in various tumor developmental processes, which we will discuss below. Here, we will explore the effects of TGF-β on different aspects of tumor growth.

3.2.3 Immune Evasion

Since sustained immune system activation can induce inflammation and tissue damage, the immune system is often rewired to prevent excessive activation. Tumor cells and other cells in the TME take advantage of these immunological safeguards by overproducing immunosuppressive cytokines, particularly TGF-β. By inhibiting various components of both innate and adaptive immunity, TGF-β creates a favorable environment for tumor growth (Figure 4). Hence, TGF-β signaling is an essential part of tumor immunosuppression.

3.3 TGF-β-Driven Immune Tolerance in T Cells

In the adaptive arm of the immune response, TGF-β activation in the TME is associated with significant suppression of T-cell differentiation, activation, proliferation, and cytokine production, leading to loss of T-cell effector function (Figure 4). This T cell-dependent anti-tumor inhibition occurs via various mechanisms. In an In Vitro study evaluating the co-culture of naïve T cells with TGF-β, it was found that naïve T cells could not differentiate into a Th1 response. In contrast, mice lacking TGF-βR2 exhibited a significant Th1 response and increased sensitivity to TCR stimulation [22, 110]. Mechanistically, the absence of TGF-βR2 in CD4+ CD8+ T cells results in enhanced TCR-dependent activation in response to weak antigenic stimulation, leading to improved functionality characterized by increased expression of IFN-γ and granzyme B [22, 110] that infiltrates into the tumor beds and tumor-draining lymph nodes [tDLN], indicating a role of TGF-β in preventing tumor-specific T cell priming. Transgenic mice with dominant-negative TGF-βR2 in CD4+ and CD8+ T cells rejected cancer cells upon inoculation, with expansion of tumor-specific CD8+ T cells [22].

Additionally, TGF-β silences critical cytokines such as IL-2, which induces T cell proliferation, interferon-γ (IFNγ), and cytolytic molecules like granzymes [22, 111]. This also inhibits the expression of T-Box expressed in T cells (T-BET) and eosodermin (EOMES), which support effector and memory T cell function [79, 112]. Activating the nuclear factor of activated T cells (NFAT) is vital for differentiating helper T cells, particularly Th1 and Th2 cells, via the interleukin-2 inducible T cell kinase (ITK). TGF-β can inhibit ITK phosphorylation, reducing calcium influx and impairing NFAT activation, contributing to reduced T cell differentiation and proliferation [113]. Furthermore, TGF-β can also downregulate the expression of key transcription factors critical for the development of Th1 and Th2 cells, such as GATA-binding protein-3 (GATA-3) and T-Box expressed in T cells (TBET), in CD4+ T cells [22].

Moreover, it is well established that prolonged antigen stimulation of cytotoxic T lymphocytes (CTLs) results in exhaustion of T cell function, marked by high expression of inhibitory receptors, including programmed cell death protein 1 (PD-1). TGF-β transcriptionally enhances PD-1 expression by recruiting SMAD3 to the programmed cell death protein 1 (PDCD1) promoter region [114]. In another instance, TGF-β-mediated T memory cell exhaustion is due to the upregulation of CD70 on effector T cells [115].

The induction of forkhead box P3 (FOXP3) by TGF-β can promote the development of CD4+CD25+ regulatory T cells (Tregs), a known mechanism of T cell immunosuppression [22]. In a seminal study by Yang and colleagues [116], TGF-β secreted by lymphoma cells promoted regulatory T (T_reg) cells’ function, mediated by the enhanced FOXP3.

3.4 TGF-β as a Modulator of B Cell Immunosuppression

The humoral response of B cells is driven by the secretion of immunoglobulins (Igs). For this process to take place, a transition from membrane-bound Igs to soluble Igs is essential. This switching mechanism can be inhibited by TGF-β, which selectively suppresses immunoglobulin lambda (Ig λ) light chains and the isotypes immunoglobulin M (IgM) and immunoglobulin G (IgG) while preventing class switching to immunoglobulin A (IgA). However, the IgA induced by TGF-β is insufficient to support robust immune responses, such as antibody-dependent cellular cytotoxicity and antibody-dependent cellular phagocytosis (ADCP) [3]. Mechanistically, TGF-β1 and TGF-β3 hinder key molecules involved in cell cycle division and transcription factors critical for the differentiation into antibody-secreting cells, such as interferon regulatory factor 4 (IRF4), B-lymphocyte induced maturation protein-1 (Blimp-1), and x-box binding protein-1 (XBP1) [117]. Additionally, TGF-β downregulates the secretion of the stromal factor IL-7, which is vital for the proliferation and survival of early lymphoid progenitors of B cell precursors. Furthermore, TGF-β reprograms the balance of pro-apoptotic and anti-apoptotic genes by upregulating BIK and downregulating BCL-XL, ultimately leading to B cell apoptosis [111].

3.5 TGF-β as a Suppressor of Innate Myeloid Function: Antigen-Presenting Cells (APCs), Macrophages, Neutrophils

In addition, TGF-β signaling suppresses the innate immune response against tumors, facilitating their progression. Studies have shown that it inhibits cell death mediated by macrophages and monocytes by promoting a shift in macrophage polarization from pro- to anti-inflammatory (M1 to M2), which reduces their effectiveness and increases the release of anti-inflammatory cytokines, thus worsening the immunosuppressive environment of the tumor [118, 119]. TGF-β also diminishes the anti-tumor activity of neutrophils as blockage of TGF-β leads to recruitment of neutrophil-attracting chemokines C-X-C motif chemokines (CXCL12 and CXCL5) and cytokines (TNF-α and IFNγ) that expresses a high level of lymphocyte antigen 6 complex locus G (CD11b+/Ly6G+) neutrophils [120]. In colorectal cancer (CRC), TGF-β mediates the polarization of tumor-associated neutrophil (TAN) from pro-tumor (N2) to anti-tumor (N1) via PI3K/AKT signaling on TANs and TGF-β/SMAD signaling in the tumor cells [121]. Additionally, on dendritic cells (DCs), TGF-β suppresses the expression of major histocompatibility complex II (MHCII), hindering antigen recognition [122].

In the TME, TGF-β-mediated immunosuppression can also be facilitated through paracrine or autocrine induction of CAF [123] and activation of MMPs [22]. The primary mechanism through which TGF-β facilitates immune exclusion is via CAF signaling, which leads to increased ECM deposition within the TME, subsequently impeding the infiltration of CD8+ T cells. In a sophisticated study, a class of ECM genes called cancer-associated ECM (C-ECM) has been identified [124]. These genes, present in hot and cold tumors, were significantly more abundant in ICB non-responders than in responders [124]. In addition, TGF-β-induced CAFs can exert inhibitory effects on anti-tumor immune cells through the polarization of anti-tumor myeloid and NK cells into a pro-tumor phenotype. This process may also occur via the secretion of physical barrier proteins, such as fibrils in the ECM, which have the potential to impede the infiltration of cytotoxic T cells [125, 126]. Taken together, these observations indicate that TGF-β signaling undermines the immune function of various cell types crucial to the body's anti-tumor defense system, including natural killer (NK) cells, CD4+ and CD8+ T cells, macrophages, neutrophils, and B cells [59, 127].

Additionally, the increased secretion of factors such as indoleamine 2,3-dioxygenase (IDO), interleukin-10 (IL-10), and arginase can redirect DCs toward an immature phenotype characterized by immunosuppressive features [128]. In melanoma, breast cancers, and other solid tumors, TGF-β signaling upregulates IDO in plasmacytoid DCs and CCL22 in myeloid DCs. This DC phenotype enhances the migration of regulatory T cells (Tregs) and contributes to overall immunosuppression [129].

Another crucial aspect of the innate immune system is the cyclic GMP-AMP synthase (cGAS-STING) pathway. Activating this pathway produces Type I IFNs and various inflammatory cytokines, significantly enhancing anti-tumor immunity [130, 131]. The cGAS-STING pathway is acknowledged as a tumor-suppressive mechanism that can facilitate the maturation and mobilization of APCs [131, 132]. When its activation is altered, it results in increased tumor survival and metastasis. For example, in multiple myeloma (MM), Tregs-induced TGF-β1 secretion was shown to inhibit the cGAS-STING signaling pathway, suppressed major histocompatibility complex (MHC-I) and upregulate programmed death ligand 1 (PD-L1), thereby contributing to a tumor-progressive microenvironment [133]. Interestingly, either neutralizing TGF-β1 or activating cGAS-STING pathway rescued the effect of Tregs on the tumor antigen presentation and immunosuppressive signal in MM cells, an effect that was faithfully recapitulated in the mouse model. This model highlights the critical role of the Treg-TGF-β1-cGAS-STING axis in mediating TGF-β1-driven immunosuppression, particularly in MM. Furthermore, accumulation of TGF-β in the TME of advanced tumors has been found to shift the balance of cGAS-STING from tumor surveillance to tumor progression by downregulating cGAS-STING levels in gamma-delta (γδ) T cells, aiding immune evasion [134].

3.6 TGF-β as a Negative Regulator of NK Cells

NK cells possess a remarkable ability to kill, making them a powerhouse of the innate immune system. However, TGF-β signaling diminishes the cytotoxic capacity of NK cells by downregulating NK group-2 member D (NKG2D) and NK cell protein 30 (NKp30), two critical activating surface receptors involved in eliminating cancer cells [135]. Furthermore, the inhibition of E4-promoter binding protein-4 (E4PBP4) by SMAD3 can further impede the development of NK cells [136]. Additionally, the negative regulation of cytolytic proteins such as granzymes and perforin via SMAD signaling can also compromise the cytotoxic potential of NK cells [137]. Another notable observation is how tumor cells exploit the plasticity of NK cells. Analysis of tumor samples revealed that the presence of TGF-β in the TME encourages the transdifferentiation of NK cells into innate lymphoid cells Type-1 (ILC1), which lack cytotoxic function, in a SMAD4-independent TGF-β signaling pathway [138, 139].

4.1 Small Molecule Kinase Inhibitors

The binding of the TGF-β receptor to its receptors (TGF-β1, 2, 3) initiates a cascade of phosphorylating events leading to the activation of SMAD2/3 and SMAD4 [140]. The phosphorylated SMADs then form a “SMAD complex” and translocate into the nucleus to increase the expression of genes that regulate tumor growth and division. Galunisertib (LY2157299) is a small molecule experimental inhibitor of TGF-β receptor kinase Type 1 that prevents the binding of adenosine triphosphate (ATP) to TGF-β thus, blocking the downstream activation of SMAD2/3 complexes and activation of TGF-β [141]. As demonstrated in the Phase 1 trial, the inhibition of these complexes prevents the development of lung and breast cancer with a safe therapeutic profile in various solid tumors [142]. It has a selective capacity to inhibit the phosphorylation of serine/threonine residue, thereby preventing the downstream “SMAD complex” formation. In addition to its primary function, these inhibitors also act on other TGF-β-related protein targets such as p-P38 MAPK, PI3K/AKT, nodal, activin, and myostatin [142, 143], endowing them with pleiotropic activity against both canonical and non-canonical TGF-β signaling. However, the safety and pharmacokinetics were evaluated in a Phase 1 clinical trial of Galunisertib with Durvalumab, an anti-PD-L1. The combined treatment failed to elicit a superior efficacy over individual drugs. The median overall survival (OS) and progression-free survival (PFS) were 5.72 months and 1.87 months, respectively [144]. A non-randomized Phase II/III study of LY2157299, in patients with very low, low, medium, and intermediate risk-myelodysplastic syndromes (MDS) showed an acceptable pharmacological and safety profile with a linked hematological recovery (24.4%) in patients with lower and intermediate-risk MDS, as well as responses in patients that depend hugely on transfusion and those with evidence of early stem cell differentiation block [145]. In addition, 43.9% attained erythroid response, and 44% had a consequential debility reduction.

The therapeutic activity of Galunisertib (NCT01246986) was examined in second-line hepatocellular carcinoma (HCC) patients in baseline serum alpha-fetoprotein (AFP). Median time-to-progression (TTP) and OS were 2.7 months, 7.3 months (n = 102), and 4.2 months, 16.8 months (n = 40), respectively, majorly observed in patients with a normal fetoprotein and reduced TGF-β1. The commonly seen treatment-related adverse effects were increased bilirubin, hypoalbuminemia, embolism, anemia, fatigue, and neutropenia. Overall, treatment with Galunisertib displayed a decent safety profile and more prolonged survival in AFP and TGF-β1 responders [146]. A Phase II research combination of Galunisertib and Lomustine (NCT01582269) in patients with recurrent glioma failed to elicit a better OS than placebo plus Lomustine [147]. In a Phase II trial of resected pancreatic cancer patients, a combination of Galunisertib and gemcitabine resulted in a boost in OS in patients treated with the combination (10.9) as against those treated with Gemcitabine and placebo (7.2 months) [144]. Another Phase 1b dose-finding/escalating study evaluated the safety, tolerability, and efficacy of Galunisertib (LY2157299) in combination with Durvalumab, an anti-PD-L1 in recurrent/refractory metastatic pancreatic cancer, displayed a limited therapeutic efficacy, with OS of and PFS of 5.72 and 1.87 months, respectively. The combination pharmacokinetic profile commensurate to the individual drugs and lack of association between probable biomarkers and treatment outcomes suggests treatment regimen as an earlier treatment line and in patients with predictive TGF-β1 inhibition [144].

Another possible anti-cancer agent, Vascosertib, is an orally available inhibitor of activin-receptor-like kinase 5 (ALK5) that has been implicated in aiding many tumor growths [148]. By inhibiting TGF-βR1 overexpression, tumor development slows down, and the host's immune response is restored. With IC50 of 17.3 nM, Vactosertib inhibits ALK-4 in different mouse models, including osteosarcoma [149-153]. Through tumor intrinsic and extrinsic mechanisms, Vactosertib exhibited potent anti-tumor effects in various solid tumors, including MM [154]. Numerous completed studies have reported on the safety and effectiveness of Vactosertib against different cancer types delineated in the table (Table 1). In a Phase I first-in-human dose escalation of Vactosertib in subjects with advanced malignant tumors, the safety, tolerability, pharmacokinetics, and an expansion phase to determine the recommended dosing for Phase II was carried out. Due to its short half-life, multiple once-per-day dosing was postulated [155]. Several ongoing and completed trials are now in the pipeline to evaluate the efficacy of Vactosertib as a combination therapy. Vactosertib, in combination with pembrolizumab, underwent an open-label multi-center trial to evaluate its safety, tolerability, and anti-cancer effect in patients with microsatellite-stable metastatic CRC [156].

4.2 Targeting TGF-β With Neutralizing Antibodies

The multifunctional role of TGF-β in both tumorigenesis and non-tumorigenesis status makes it an attractive target in cancer. However, it could also be a source of systemic toxicity, as confirmed by several studies [157]. On this note, it is paramount to understand and evaluate the role of the three isoforms in tumor progression, serving as a basis for selective toxicity. Although the dominant form in tumors is TGF-β1, the role of other isoforms is yet to be fully elucidated. In two pre-clinical models evaluating the blockade and efficacy of TGF-β1 or TGF-β3 isoforms in colon and melanoma cancer, neutralization of TGF-β1, but not TGF-β3, increased the prophylactic efficacy against a colon cancer vaccine, promoting tumor-infiltrating CD8+ T cell, and reducing regulatory T-cells and myeloid-derived suppressor cells (MDSCs) [158]. In this same study, TGF-β1 isoform blockade, as a single agent, improved mouse survival by delaying tumor intrinsic phenotypic EMT [158]. This evaluated the critical role of TGF-β1 against other isoforms and its therapeutic potential either as a monotherapy or combined with other immunotherapeutic or targeted therapies. In another randomized tamoxifen trial study of 564 invasive pre-menopausal breast cancer carcinoma, TGF-β2 expression in CAF was associated with improved recurrence-free survival (RFS), suggesting TGF-β2 could have an anti-tumorigenic effect in CAF⁺ breast cancer patient, thereby serving as a novel prognostic marker, especially in transformed-CAF+ cancer [159]. It is, however, not surprising that the blockade of TGF-β1,2,3 to TGF-βR2 attenuated the growth and metastasis of cancer stroma, suggesting a conflicting dual role of these isoforms in different cancer models. This further supports the need for extensive studies of the role of the three isoforms in cancer and characterizing the TME as part of diagnostic tools in cancer patients. This could inform if a patient will benefit from TGF-β inhibitors.

In the Phase 1 clinical trial (NCT00356460) of Fresolimumab in advanced and malignant melanoma or renal carcinoma, Fresolimumab, a human IgG4 neutralizing monoclonal antibody against the three TGF-β isoforms, demonstrated a tolerable safety profile [160]. However, this trial was conducted in a small cohort of 29 patients and may not accurately represent this therapy's safety profile. In a discontinued Phase 2 trial of the immunological effect of Fresolimumab in malignant pleural mesothelioma, 3 out of 13 patients recruited showed stable disease with improved OS but no impact in effector adaptive immune cells [122].

During a Phase I clinical trial of anti-TGF-βII IgG1 monoclonal antibody in patients with advanced tumors (NCT01646203), a treatment aimed at inhibiting receptor-mediated signaling activation of TGF-β was evaluated in 14 patients. The study revealed various treatment-related adverse events, ranging from uncontrolled cytokine release syndrome to infusion-related reactions. Although three dose ranges were tested, they failed to determine the maximum tolerable dose (MTD). Furthermore, escalating the dose beyond 25 mg/dose was deemed unsafe due to the observed adverse events [161]. Other neutralizing antibodies were tested in preclinical. These include 2G7, an antibody that inhibits the three TGF-β isoforms in nude mice of human breast cancer and metastatic lung cancer [162]. In addition, 2G7 showed increased NK cell activity implicated in the anti-tumor effects. 1D11 is another anti-TGF-β monoclonal antibody that inhibits the three isoforms and was investigated in preclinical models of lung and breast cancers. However, at higher doses, 1D11 induced epithelial hyperplasia of the tongue in mice, observed with extreme weight loss and dysphagia in a mouse model of familial adenomatous polyposis, 1D11 increased cancer progression. Hence, it is necessary to determine if blockade of TGF-β could induce tumor progression, especially in genetically related cancers [163].

4.3 Ligand Trap

The constitutive ligand-receptor interaction of TGF-β that follows when the cytokine binds to its surface receptor, causing dimerization and further downstream signal transduction, can be further inhibited with either anti-ligand or anti-receptor antibodies or soluble TGF-β receptor inhibitors known as ligand traps or sequesters [164]. Most of these inhibitors are designed to isolate TGF-β1 and TGF-β3 receptors by preventing their binding and activation to TGF-β2 [164]. Although these antibodies displayed remarkable pharmacokinetic parameters via high binding capacity and superior specificity to the receptors, they have less capacity to efficiently infiltrate and distribute evenly in tumor mass [3, 4]. So far, the primary strategies employed in developing ligand trap sequesters include anti-ligand/receptor. The first soluble TGF-βII trap was developed in a transgenic metastatic breast cancer mice model [165]. The antagonist was able to confer lifetime protection by preventing metastases in mice when rechallenged with orthotopic tumor implantation [165]. A novel tetravalent TGF-βII trap showed comparable inhibition to a pan TGF-β kinase inhibitor or neutralizing antibody in different prostate cancer cells [166]. Since then, several other antagonist and/or ligand traps have been developed, although with poor activity, and some are already in clinical trials (Table 1).

4.4 Antibodies Against Integrin-Mediated TGF-β Signaling

Recruitment of cell adhesion molecules to ECMs is vital to regulating homeostasis. Integrins are the central cell adhesion receptors and migratory molecules crucial in cell movement [196, 197]. Altered integrins have been implicated in almost all cancers, and they not only contribute to cancer progression but play a significant role in tumor metastasis [196, 198-200]. They have also been linked to aberrant oncogenic signaling, growth factor signaling (GFR), T cell resistance, colocalization, and anchorage-independent survival of tumors in metastatic sites [201, 202]. Additionally, integrins could remodel the tumor stroma, which supports the invasion, acquisition, and maintenance of cancer stem cells (CSCs), ultimately leading to resistance to therapy [203]. There is increasing evidence that integrins play a role in activating TGF-β1 ligand [204]. When integrins bind to TGF-β1, they can negatively impact the tumor stroma, which can cause the tumor to progress [205]. Therefore, targeting integrins could be a viable strategy for blocking TGF-β signaling. When TGF-β1 interacts with β, av-integrin complexes, it activates latent TGF-β signaling. Notable integrins that activate TGF-β include αvβ1 (40), αvβ3 [206], αvβ5 [206], αvβ6 [207], and αvβ8 [208]. Each of these integrins has specific expressions in different tissues. By disrupting this interaction, it may be possible to provide targeted, tissue-specific therapy, preventing systemic inhibition, which could be harmful, as TGF-β has pleiotropic effects, thus, monoclonal antibodies against αvβ1 (44,45), αvβ6 [209], αvβ8 [210, 211] have shown optimal inhibition of TGF-β1. Few preclinical studies evaluating the blocking of these interactions have been reported with αvβ8 in colon carcinoma and breast cancers [212], where the blockade of αvβ8 using engineered antibodies blocks the release of TGF-β1, mediating an anti-tumor immune effect. A neutralizing antibody, 264RAD, against αvβ6 also inhibits TGF-β1 signaling and suppresses the tumor growth in αvβ6+ human PDAC xenografts [213].

Another antibody, SRK-181-mIgG1, has been developed to target the TGF-β large N-terminal of the precursor polypeptide, also known as latency-associated peptide (LAP). This prevents the detachment of mature TGF-β1 from latency-associated proteins, keeping TGF-β1 latent and averting the activation of TGF-β2 and TGF-β3. This, in turn, prevents systemic activation of TGF-β. The antibody was tested for its ability to prevent checkpoint blockade resistance [214], thus preventing systemic activation of TGF-β. This agent, tested as a monotherapy or combined with anti-PD1 blockade, produced a robust anti-tumor effect via CD8+ T cells and decreased immunosuppressive myeloid population. Providing a rationale for selective blockade of TGF-β [215]. In addition, a clinical trial evaluation (NCT04152018) of PF-06940434 that targets αvβ8 is currently ongoing.

4.5 Vaccine-Based Treatment Therapy

Gene-based vaccine therapy is a notable strategy against TGF-β. Belagenpumatucel-L (Lucanix) is a non-viral allogeneic tumor-cell vaccine containing four lung cancer cell lines transfected with a TGF-β2 antisense transgene that downregulates TGF-β2 signaling [216]. In Phase II studies in 83 and 75 NSCLC patients of various stages, Belagenpumatucel-L showed improved survival, especially in patients with advanced-stage NSCLC [217]. This prompted a Phase III randomized trial, where Belagenpumatucel-L was given intradermally to 270 patients. However, the study failed to reach a significant difference in OS and PFS (median OS: 20.3 months) to placebo (17.8 months). Although, the study showed that early treatment in patients who received chemotherapy and radiation may be associated with better benefits, suggesting that the vaccine may be beneficial in this category of patients [218].

Using a principle similar to Belagenpumatucel-L, GVAX was developed to deliver the whole tumor cell vaccine and granulocyte-monocyte colony-stimulating factor (GM-CSF). Combined with cyclophosphamide (GVAX-Cy), this vaccine was tested in 20 patients with stage IIB-IV melanoma, with patients receiving four doses over 28 days. Elevated levels of activated monocytes, PD-1+ lymphocytes, and eosinophils were observed. However, there was no decrease in Tregs level [219]. Several other Phase I/II studies in NSCLC patients displayed promising benefits with increased granulocytes, DCs, lymphocytes, and macrophages [218, 220]. However, the treatment failed to elicit a similar effect in two Phase III trials with prostate cancer patients [221, 222]. In another Phase II study, with pembrolizumab in patients with mismatch repair-proficient (MMR-p) advanced CRC, ORR was measured. The study failed to meet its ORR, with PFS and OS of 82 days and 179–441 days, respectively. However, some patients exhibited an increased immune response [223]. Another Phase II study evaluated GVAX with ipilimumab as a maintenance therapy for metastatic pancreatic cancer. The study failed to improve the overall response or survival of metastatic pancreatic cancer as maintenance therapy [224].

Another notable vaccine is Gemogenovatucel (Vigil), which is derived from the patient's tumor cells during the initial de-bulking. Vigil is engineered with a plasmid that encodes GM-CSF and a furin short-hairpin (shRNA) that knocks down furin gene expression. GM-CSF acts to promote antigen immunogenicity and presentation by DCs, and furin is a crucial convertase needed for the secretion of TGF-β1 and TGF-β2. Vigil has been tested in several trials such as a Phase II/III multicenter randomized double-blinded trial of 80 patients with different high-risk stages (3b-4) of ovarian cancer [225], a Phase II cross-over trial of 42 cancer patients that tested for OS and RFS [226]. In a separate Phase II study, the safety and efficacy of Vigil in women with advanced gynecological cancers were examined, both as a standalone treatment and in conjunction with checkpoint inhibitors like Atezolimumab [226]. This ongoing study aims to gauge biomarkers of response, including anti-tumor immune response. Another preliminary study assessed the safety, tolerance, and efficacy of Vigil, either alone or combined with Durvalumab, and measured PFS in the combination arm for patients with advanced gynecological cancers [226]. Overall, Vigil was well tolerated and met some of the endpoints such as OS and RFS.

4.6 Anti-Sense Oligonucleotides (ASOs) Therapy

Oligonucleotides are a novel class of agents used to treat various diseases, including cancer. However, they are highly unstable, and due to this, the modified version of ASOs using nanoparticles or chemical modification has been developed to aid delivery and stability [227]. In a few preclinical studies using AP11014 and AP15012 oligonucleotides in mouse models of prostate cancer, CRC, non-small cell lung cancer (NSCLC), and MM, suppression of TGF-β was adequately achieved [228, 229]. Additionally, AP12009 (Trabedersen), which inhibits TGF-β2, demonstrated the ability to impede cell proliferation and migration in a preclinical study involving human pancreatic cells and an orthotopic xenograft model. Consequently, a Phase I/II trial was conducted to assess the safety and efficacy of Trabedersen in various cancers, including pancreatic cancer, malignant melanoma, and CRC [230]. The prosecution revealed favorable safety and survival outcomes, with the primary adverse reaction being non-serious, short-term thrombocytopenia [231]. Furthermore, AP12009 was studied in an open-label Phase I/II dose-escalation trial in recurrent high-grade glioma, with a satisfactory median survival profile [230].

4.7 Combination-Based Therapy

Advances in cancer immunotherapy, specifically ICIs and chimeric antigen receptor T cell therapies (CAR-T) have transformed cancer treatment since discovering PD-1. These agents are classified into two categories: co-stimulatory agonists such as CD26, CD28, CD40, OX40, inducible costimulatory (ICOS) CD134, and CD137, and co-inhibitory antagonists like PD-1, cytotoxic T-lymphocyte associated antigen-4 (CTLA4), lymphocyte activation gene 3 (LAG3), T cell immunoglobulin and mucin domain 3 (TIM3), and T cell immunoreceptor with Ig and ITIM domain (TIGIT). However, these therapies have recorded a sub-optimal response rate of only 15% in patients, partly due to TGF-β-immunosuppression [232]. TGF-β has been found to contribute to anti-PD-1/PD-L1 treatment resistance by promoting tumor evasion and downregulating key inflammatory cytokines and cytolytic proteins such as granzyme A, B, perforin, IFN-γ, and Fas ligand via SMAD-mediated mechanisms that attenuate CD4+ and CD8+ T cell responses [170, 233, 234]. Furthermore, immune-excluded tumors exhibit significant elevation in TGF-β signaling in the stromal compartment, which might play a role in preventing immune infiltration. Therefore, inhibiting TGF-β could suppress immunosuppressive factors that promote angiogenesis and EMT [233-235]. Given the similar roles of PD1/L1, CTLA4, and TGF-β in the immunosuppression of T cells, macrophages, and NK-cells, treatment strategies that combine ICIs with TGF-β inhibition are more likely to improve therapeutic outcomes.

Additionally, there is growing evidence that both TGF-β and PD-L1 are parallel immunosuppressive pathways that can activate each other, as observed in some tumor cells where TGFβ-induced EMT increases the expression of PD-L1 and TGF-β-SMAD3 signaling enhances PD-1 levels on TILs [114, 236, 237]. Therefore, there is growing interest in combining anti-TGF-β and ICIs. Moreover, targeting TGF-β in the TME before or after adoptive cell therapy leukapheresis might significantly reduce T cell immunosuppression [238]. Based on this, several preclinical combination studies are being carried out, with targets demonstrating more significant superiority when combined with anti-TGF-β. In a mouse xenograft model of triple-negative breast cancer (TNBC) and melanoma, a combined effect of TβRII chimera with anti-CTLA4 demonstrated profound anti-tumor responses compared to anti-CTLA4 alone [177]. The same is true of a bispecific decoy of CD4+ TGF-β trap, which blocks TGF-β signaling on CD4+ T cells, leading to tumor vasculature normalization in a transgenic model of breast cancer [239]. In a mouse breast cancer model of Galunisertib (LY2157299) with anti-PD-1/L1, the combined treatment displayed superior anti-tumor efficacy than the single arm [143]. Other similar combinations of preclinical models have been tested with most currently in clinical trials, such as AVID200 in homografts tumor models [240], SRK-181-mIgG1 with anti-PD1 in ICI-resistant syngeneic mouse models showed a superior infiltration of CD8+ T cells [215], a decoy of CD4+ TGF-β trap with a selective blockade of TGF-β signaling in CD4+ T cells and reduced immunosuppressive myeloid cells [241], and pH-sensitive nanoparticles for Galunisertib (LY2157299), with small interfering PD-L1 RNA have been used to suppress tumor growth in PDAC mouse models [242], and finally is TβR1 inhibitor SM16A, which enhances adoptive T cells infiltration and anti-tumor effect in mesothelioma mouse models [243]. Shou et al. have recently developed a hemagglutinin-tagged decoy comprising the TGF-βRII ectodomain and a glycophosphatidylinositol (GPI) anchor. This construct demonstrated elevated expression of IFN-γ and granulocyte-macrophage colony-stimulating factor (GM-CSF) alongside other cytokines. It exhibited proficient capability in trapping and deactivating TGF-β [244]. Numerous studies have explored the combination of TGF-β inhibitors with other cancer therapies. Some of these investigations have progressed to clinical trials, as delineated in Table 1, and have exhibited manageable safety, tolerability, and anti-tumor efficacy. However, none of these interventions have received approval for clinical use. A promising strategy for solid tumors is CAR-T cell therapy, which entails the modification of T cells to target tumor antigens. While CAR-T cell therapy has exhibited considerable success in hematologic malignancies, its efficacy against solid tumors has been limited, partly attributed to TGF-β-mediated immunosuppression [245-247]. Researchers have developed strategies to overcome this hurdle, including co-expressing the negative dominant form of TGF-β (TβII) [248] and the constitutively active AKT (caAKT) [249], which provides resistance against tumor-associated inhibitory signals. Additionally, co-expressing the TβIII-41BB chain has prevented immunosuppression by converting the TGF-β immunosuppressive signal to a 41BB stimulatory signal [250].

Our analysis of nearly two decades of anti-TGF-β treatment development highlights the need to review unsuccessful clinical trials for insights. This evaluation is crucial for future therapeutic advancements. Nisevokitug (NIS793), a human IgG2 monoclonal antibody aimed at pancreatic adenocarcinoma (PDAC), showed inadequate tumor shrinkage, leading to its discontinuation (NCT04935359). Additionally, a Phase II trial of Bintrafusp Alfa, intended for metastatic biliary tract and lung cancers, did not meet its primary endpoint compared to Nivolumab, showing a lower ORR and more severe adverse events (NCT03838661) [1]. Integrin inhibitors also showed limited efficacy: an antibody targeting integrins αvβ3 and αvβ5 in glioblastoma (NCT00689221) and head and neck SCC (NCT00705016) did not improve survival or anti-tumor efficacy compared to chemotherapy. Similarly, a trial for patients with advanced solid tumors failed due to severe cytokine release syndrome and challenges in determining the MTD (LY3022859) (NCT01646203). As previously articulated, the favorable outcomes from numerous clinical trials involving anti-TGF-β therapies have yet to be realized. Even in instances where noteworthy results have been documented, they remain neither universal nor sufficiently satisfactory when juxtaposed with data derived from preclinical investigations. Drawing upon recent discoveries and an exhaustive review of the literature, we posit several significant considerations regarding TGF-β signaling that are frequently overlooked in the majority of clinical assessments. In particular, we underscore: (i) the off-target effects, (ii) the limitations of preclinical models, (iii) the activation of redundant and compensatory mechanisms, (iv) the absence of patient selection criteria, and (v) the dual opposing roles of TGF-β at both contextual and cellular levels.

5.1 Off-Target Effects

Compelling evidence suggests that cancer patients treated with TGF-β blockers may experience significant organ toxicity. This is because TGF-β is ubiquitously expressed in nearly all healthy cell types where it regulates critical homeostatic functions [251, 252]. For instance, while TβRI kinase inhibitors showed therapeutic effects in cancer patients, there was significant cardiac toxicity at high doses. These include hemorrhagic, degenerative, and inflammatory lesions in heart valves. Additionally, skin toxicity manifested as eruptive keratoacanthomas, hyperkeratosis, cutaneous squamous-cell carcinomas, and basal cell carcinoma, limiting the safe therapeutic window of these inhibitors [253]. Addressing these challenges may require a focused strategy to identify and target definite genes modulating the tumorigenic activity of TGF-β, such as targeting specific epitopes of TGF-β on cancer cells and other relevant cell types to minimize off-target toxicity while preserving normal homeostatic function. In addition, implementing an intermittent dosing schedule, where patients experience planned breaks from the medication (called drug holidays), may be crucial for Patients' safety. For instance, the treatment protocol for alisertib in patients with glioma in 2 weeks on and 2 weeks off schedules significantly reduced cardiac toxicity [254]. In addition, the timing of anti-TGF-β therapy is crucial for optimal outcomes. Patients with premalignant or early-stage malignant tumors may derive more significant benefits from this therapy, as the tumor-suppressive capacity of TGF-β remains active during these stages [71]. In contrast, advanced tumors may have developed uncontrollable signaling networks, rendering them less responsive to anti-TGF-β treatment. This rationale is evidenced in mouse and animal studies where the TGF-β was inhibited before exposing the mouse to tumors. The study observed a significant decrease and, in some cases, complete eradication of tumor growth with high infiltration of CD8+ T cells [255]. Given the safety of plant-derived compounds and their potent anticancer activities [256, 257], exploring naturally derived compounds as inhibitors of TGF-β in cancer therapies may be rewarding. Interestingly, a plant-derived agent such as 8-Oxo-epi berberine (OPB) was shown to reverse TGF-β1-triggered morphological changes, EMT biomarkers expression, migration, adhesion, and invasion in cell culture assays [258]. In animal models, OPB also dramatically decreased the metastatic nodules in the lung without affecting the growth of primary tumors. In conclusion, OPB inhibited TGF-β1-induced EMT, possibly by interfering with SMAD3. This study further revealed that OPB suppressed TGF-β1-induced SMAD2/3 activation, SMAD3 phosphorylation and nuclear translocation, and interaction of SMAD3 with SMAD4.

5.2 Limitations of Preclinical Testing Models

Another potential factor responsible for the poor clinical performance of TGF-β inhibitors could be the misrepresentation of preclinical data that do not faithfully recapitulate the complexity of the disease in humans. For instance, Maier et al investigated the effectiveness of Galunisertib in patient-derived tumor xenografts (PDXs) and found impressive tumor inhibition in lung and prostate cancer PDX models [259]. However, this agent failed to demonstrate improved OS compared to placebo in clinical studies (104). Despite providing a crucial platform for preclinical research and personalized medicine, PDXs have significant limitations. First, the fidelity of cancer cells in the PDX models is compromised due to loss of heterogeneity, clonal evolution within the tumors, and selection bias for engraftment. A study analyzed 1110 PDX samples across 24 cancer types to understand the dynamics of copy number alterations (CNAs) dynamics during PDX passaging [260]. The findings revealed that CNAs rapidly accumulate during the passaging of PDXs, likely due to the selection of preexisting minor clones within the tumors. Notably, the accumulation of CNAs in PDXs differed from that acquired during tumor evolution in patients. Notably, several CNAs recurrently observed in primary tumors tended to disappear over time in PDXs. These findings suggest that the genomic aberrations in PDXs are dynamic and do not necessarily capture the genomic landscape of primary tumors. This raises a big concern about the preclinical data generated from PDX models that predict the clinical efficacy of Galunisertib. Another critical limitation of PDX models that may explain why Galunisertib was ineffective in humans despite encouraging results in PDX models is the lack of immune cells in the TME compared to human tumors. This means that the efficacy of Galunisertib as reported in PDX models disregards the contribution of the immune components in tumor progression. Stressing the critical role of the TME in TGF-β signaling, it was found that the cytotoxic activities of Galunisertib are dependent on factors emanating from the TME [259]. Humanized mouse models offer a more sophisticated platform for anticancer drug assessment, closely simulating human disease conditions. These advanced models, which incorporate human immune system components into mice, offer improved predictive capabilities for drug screening within a human-like immune environment. This innovative method signifies a significant leap forward in preclinical research, generating findings that more accurately forecast potential outcomes in human patients.

5.3 Activation of Redundant and Compensatory Mechanisms

Clinical insights suggest that one of the key resistance mechanisms to anticancer therapy is utilizing functionally redundant pathways or activating compensatory mechanisms, leading to adaptive and intrinsic drug resistance. The complexity of the TGF-β signaling pathway may promote redundancy [261]; for instance, five TGF-β receptors, two TβR1 (ACVRI, BMPIa), and three BMP-SMAD receptors (SMAD1, SMAD5, SMAD8) can redundantly transduce anti-mullerian hormone signal needed for male infertility [262]. Also, reports have shown that chemoresistance is a major clog in the wheel of TGF-β inhibitors for cancer therapy [263]. While the molecular mechanisms underlying this resistance are not fully clear, it is tempting to speculate that activation of compensatory mechanisms would be a major culprit. This can occur at receptor or pathway levels. At the receptor level, TGF-β has 14 signaling receptors, 42 ligands, and several proteases (serine protease, plasmin MMPs, plasma kallikrein, etc) that can activate it [264, 265]. Presently, clinical trials primarily target a select few components of this pathway, potentially leading to variations in patient responses. Excluding other signaling mechanisms might result in variation in anti-TGF-β responses in patients depending on ligand/receptor availability and accessibility. Therefore, comprehensive patient stratification, informed by advanced bioinformatic and biotechnological tools to profile the expression and activation of these components, is essential to discern which patients are likely to benefit from TGF-β therapy. Additionally, efforts should be geared toward identifying TGF-β redundant pathways in cancer progression that can lead to treatment failure and resistance. It might also be noteworthy to determine which type of TGF-β signaling receptors/ligands/proteases are expressed by patients and define their role in cell and tissue-dependent manner. Multiple signaling pathways can compensate for the loss of TGF-β activities at the pathway level. A prime example is observed in angiogenesis, a crucial oncogenic process that ensures continuous nutrient supply to tumor cells. While inhibiting TGF-β-mediated angiogenesis is a key objective of anti-TGF-β therapies, numerous alternative factors can effectively induce angiogenesis in its absence. These include FGFs, angiopoietins, EGFs, fibroblast growth factors, hepatocyte growth factors, placental growth factors, and stromal cell-derived factor 1 [266]. This redundancy underscores the limitations of using anti-TGF-β agents as monotherapy. Corroborating this concept, both preclinical and clinical studies have demonstrated that therapeutic strategy targeting other angiogenic factors, such as using anti-VEGFA antibodies or tyrosine-kinase inhibitors against VEGF receptors, paradoxically stimulates the production of alternative growth factors [267]. This compensatory mechanism further complicates the efficacy of single-pathway targeted treatments and highlights the need for more comprehensive therapeutic strategies. These observations justify that anti-TGF-β therapy may yield more significant benefits when combined with other treatments instead of being used as a standalone therapy. Hence, its efficacy may be enhanced when administered concomitantly with chemotherapeutic or immunotherapeutic agents. Notably, combination with immunotherapies may mitigate immunosuppression and prevent cancer invasion and metastasis.

5.4 Lack of Patients' Selection Criteria

Choosing the right treatments for an individual patient is complex but is central to achieving remission and increasing survival. Most clinical studies that investigate the effectiveness of TGF-β blockers did not have specific selection criteria, which may partly explain modest clinical benefits. As we gain a deeper understanding of the effects of TGF-β on the TME components, it becomes evident that the cellular and molecular characteristics influencing responsiveness to TGF-β blockade can vary between different tumor types and across individual patients. For instance, the response to TGF-β blockade in PDAC, which is characterized by extensive stromal development, frequent inactivation of SMAD4, and a low tumor mutation burden [268] differs from that in head and neck SCCs, which have high tumor mutation burdens and significant CD4+ T cell presence In parallel studies [269], transcriptional profiling of patient samples from many different cancer types, including glioblastoma, pancreatic cancer, breast cancer, ovarian cancer, (CRC), and NSCLC, revealed that cancer patients with mesenchymal subtypes have high expression of TGF-β target genes and that this correlates with poor prognosis. Therefore, subgroups of patients with cancers that carry a mesenchymal phenotype may particularly benefit from anti-TGF-β therapies [270-272] According to these results, the effectiveness of TGF-β blockers appears to be influenced by both the type of cancer and the stage of progression, with late-stage patients possibly seeing more significant improvements. In sum, using predictive biomarkers to determine patients who are most likely able to benefit most from treatment with anti-TGF-β agents may increase the clinical benefit of TGF-β targeted therapy. Factors such as mesenchymal gene expression signature, SMAD4 status, and tumors with extensive stromal development, such as pancreatic cancer, may be good candidates for TGF-β blockade. It should be noted that serial biopsies may be required to fine-tune treatment strategies as tumors may rewire their gene expression profile in response to treatment or with advanced tumor stage.

5.5 Dual Role of TGF-β

The dual nature of TGF-β poses significant challenges in developing therapeutic interventions, as inhibiting its tumor-promoting signaling may compromise its tumor-suppressive functions. Consequently, responses to TGF-β therapy across nearly all types of cancer are markedly context-dependent. The tumor-promoting characteristics of TGF-β have been extensively documented [16, 17, 59, 101, 149, 273-275]. Nevertheless, TGF-β exhibits tumor-suppressive activities during the early stages of tumor development; however, the extent to which this property is preserved in advanced tumors remains undetermined. Therefore, disrupting TGF-β signaling necessitates a more nuanced approach. In a controlled In Vitro study assessing the effects of TGF-β neutralizing antibodies on 12 metastatic breast cancer models, only 5 (42%) of the models exhibited a suppression of metastasis, whereas 4 (33%) displayed no response, and 3 (25%) experienced progressive stimulation of metastasis [276]. The inhibition of metastasis was found to be immune-dependent, while the stimulation of metastasis appeared to be primarily associated with tumor cells, particularly CSCs. Further transcriptomic analysis of the pre-treated tumors demonstrated differential gene enrichment; the responder group was enriched with genes related to inflammation and anti-tumor activity, characteristics of immunologically active tumors. In contrast, non-responders presented an immunologically inactive state, exhibiting heightened levels of rapamycin complex 1 (mTORC1) mammalian targets and tumor-associated macrophages (TAMs), recognized for their role in promoting immunosuppression [276].

Furthermore, the metastatic-stimulating properties of anti-TGF-β are attributed to the interference with TGF-β's tumor-suppressive effects, subsequently promoting the activities of CSCs. Notably, TGF-β inhibits CSC activity in certain melanoma and breast cancer models [277]. An additional study investigating the TGF-β tumor-suppressive gene signature indicated that the persistent expression of these genes is associated with a reduction in metastasis among breast cancer patients in this category. Collectively, this information suggests that the prevailing notion that the tumor-suppressive effects of TGF-β are diminished in advanced metastatic cancers may not be universally applicable across all patient categories within the context of breast cancer. This also implies that a subgroup of breast cancer patients with cold tumors, particularly those that are estrogen receptor-positive, may not derive benefits from anti-TGF-β therapies. The study underscores the significance of examining multiple targets rather than a singular target and the necessity of analyzing the transcriptomic profiles of cancer patients to identify candidates more likely to respond favorably to anti-TGF-β therapies. It remains to determine if a similar scenario exists in other cancer types.

Furthermore, TGF-β is recognized for its role in promoting metastasis by enhancing the activity of EMT-promoting genes. However, emerging studies from breast cancer [278], HCC [279], lung cancer [280], and CRC [281] indicate that anti-TGF-β treatment can also augment mesenchymal-epithelial transition (MET) activity, thereby enhancing colonization and metastasis, further complicating anti-TGF-β therapeutic application. Given that the majority of cancer diagnoses occur at an advanced stage when the disease has already disseminated into lymphatics and undergone EMT, the administration of anti-TGF-β therapies may hinder the maintenance of the mesenchymal phenotype while facilitating the seeding and spread of secondary tumors through MET. This observation is particularly evident in a breast cancer model exhibiting distinct phenotypes; administration of the TβR1 kinase inhibitor led to two disparate outcomes: treated and untreated epithelial-like breast cancer cells metastasized at comparable rates, whereas the mesenchymal-like breast cancer treated with anti-TGF-β exhibited an increased incidence of lung metastases, suggestive of enhanced MET activity in this context [278]. Thus, it can be inferred that TGF-β may potentially obstruct MET activity by impeding colonization in secondary organs. This further underscores the necessity of elucidating the precise role of TGF-β at the contextual level and at various cellular stages of cancer metastasis.

By and large, dual and context-dependent roles of TGF-β in cancer provides a strong rationale to consider localized blockade strategy may offer a more effective and safer therapeutic approach compared to systemic inhibition. This approach would balance efficacy with safety, ensuring that therapeutic benefits are maximized while systemic side effects are minimized. localized delivery methods, such as nanoparticle-based drug carriers, intratumoral injections, or tumor-targeted antibodies, could selectively inhibit TGF-β activity within the TME without affecting systemic signaling.